As Bregman (1990) argues, a listener's mental representation of sonic environments, or "auditory scenes," consists of multiple interacting "auditory streams." These streams form because our minds strive to group individual sounds with similar properties into coherent wholes, making it easier to assign them to specific sound sources and thus better predict the environment. Grouping encompasses both simultaneous and sequential sonic events that occur within distinct spatio-temporal hierarchies. While simultaneous grouping relates to stream formation and tends to occur within the psychological present of echoic memory, sequential grouping—involving strings of events—typically requires the more extended timespan of working memory. Thus, our perception of streams is inherently temporal: we need time to process the auditory scene, form expectations about its future evolution, and revise those expectations when they prove inaccurate. Despite this, with a few exceptions (Huron 2016; McAdams 2022), previous research has primarily focused on static snapshots of auditory environments.

My paper addresses this gap by analyzing the real-time (1) formation of streams and (2) their interaction, also considering (3) their effect on the listener's evolving sense of time. As a case study, I examine an orchestral work by Arvo Pärt, Swansong (2013). Since Pärt's compositional style, tintinnabuli, structurally induces listeners to perceive different layers of streams, his orchestral music will be ideal for exploring how streams with varying timbres and textural density affect our perception of meter.

Building on Lerdahl and Jackendoff's influential methodology (1983), I propose preference rules for (1) the formation of streams (Stream Formation Preference Rules, SFPRs), and I apply them to determine the number and composition of streams in Pärt's work. Additional preference rules (Stream Interaction Preference Rules, SIPRs) are introduced to analyze (2) the interaction of streams and determine their respective salience. Since determining the formation and interaction of streams requires the preliminary identification of a piece's auditory scenes, I also propose preference rules for their delimitation (Scene Transition Preference Rules, STPRs). Finally, I posit that the most salient stream will "emerge" in the listener's mind, shaping (3) their temporal expectations, which I illustrate using Mirka's (2009) dynamic model of meter, Jones's (2019) Dynamic Attending Theory (DAT), and my Implied Pulse Detector (IPD). By identifying the sequence of the most salient streams, I will then trace the potential fil rouge guiding the listener's real-time disentangling of polyphonic temporalities.

References

Bregman, Albert. 1990. Auditory Scene Analysis. Cambridge, MA: MIT Press.

Huron, David 2016. Voice Leading: The Science behind a Musical Art. Cambridge, MA: MIT Press. 

Jones, Mari Riess. 2019. Time Will Tell: A Theory of Dynamic Attending. New York, NY: Oxford University Press.

Lerdahl, Fred, and Ray Jackendoff. 1983. A Generative Theory of Tonal Music. Cambridge, MA: MIT Press.

Mirka, Danuta. 2009. Metric Manipulations in Haydn and Mozart: Chamber Music for Strings, 1787–1791. Oxford University Press.